Extraordinary acoustic Raman: listening to single nanoparticles

Vibrations can tell you about the size, shape, and elastic properties of the vibrating object. At the nanometer scale, they can help identify and selectively probe nanoparticles. Vibrations assist our understanding of how biomolecules such as proteins perform mechanical tasks in our bodies. We can use the vibration spectrum to diagnose diseases, and to help discover cures.

It is possible to investigate the vibrations of matter with light. For micron-scale structures in cavity optomechanics,¹ typical vibration frequencies are up to the megahertz range. Nanoparticles smaller than 50nm generally have mechanical vibrations in the so-called extremely high frequency range (> 30GHz) and even into the terahertz regime. This presents a serious challenge when trying to detect these vibrations with high-speed photodiodes. To solve this problem, there are various approaches to measure the inelastic scattering of light from material vibrations, such as Brillouin scattering, Raman scattering, and ultrafast pump-probe methods (including the optical Kerr effect).

An even greater challenge lies in probing only a single nanoparticle, since the scattering signal is very weak. Research using resonant metal nanoparticles, which act like antennas to capture the light,² has shown single-particle sensitivity. However, these demonstrations are limited to resonant nanoparticles.

We have developed a general solution to this problem that uses nanoaperture optical tweezers capable of trapping single proteins and other dielectric particles in the single digit nanometer range³ (see Figure 1). Instead of using one laser in the optical tweezer setup, we use two single-frequency lasers and tune the frequencies so that the beating intensity is modulated at the frequency difference. This modulates the electrostriction force on the nanoparticle at the beat frequency.

Figure 1. (a) Schematic of how extraordinary acoustic Raman (EAR) ‘listens’ to nanoparticle vibrations. Two beating laser beams create an electrostriction force that resonantly excites the nanoparticle at the beat frequency. This causes the nanoparticle to heat up. The resulting increase in thermal motion is detected by increased fluctuation in the laser intensity transmitted through the nanohole. By sweeping the beat frequency, we obtain the spectrum of the particle's vibration resonances. (b) Actual data showing a polystyrene nanosphere EAR spectrum and corresponding time-domain trace of photodiode voltage detecting laser intensity.

We do not attempt to measure the extremely weak scattered light directly. Rather, we use a trick well known in the optical tweezer community⁴ to measure changes in the Brownian motion amplitude of the nanoparticle to detect its temperature. When the beat frequency of the trapping lasers matches the nanoparticle's vibration resonance frequency, the nanoparticle heats up. In this way, we can find the vibrational spectrum indirectly by sweeping the beat frequency and recording the amplitude of the Brownian motion fluctuations. We call this approach extraordinary acoustic Raman (EAR).

Using this spectral technique, we measured the vibrational resonances of many different nanoparticles, and our observed vibration frequencies match well with predictions from elastic medium theory. Because our method has high resolution, we could resolve fine peak splitting in titania nanoparticles that have anisotropic (i.e., oriented) elastic properties, enabling investigation of the nanoparticles' material characteristics. In the case of titania, the difference in elastic properties along two different crystal directions splits the vibration resonance peak frequency into two different but closely spaced frequencies.

More importantly, we have used EAR to listen to biomolecules. The first four proteins we studied showed unique spectral characteristics (see Figure 2). Awareness of these would enable identification of different proteins in solution (to diagnose whether a mutant form is present, for example). Furthermore, knowledge of such low-order vibrational modes may shed light on fundamental processes in biochemistry, such as protein folding and allostery (the influence of one binding site in a protein on another distant site).

Figure 2. EAR spectra of four different proteins: (a) carbonic anhydrase, (b) conalbumin, (c) streptavidin, and (d) aprotinin.

EAR's ability to resolve protein vibrations is surprising because we generally assume that energy redistribution and viscous damping wash out the coherent vibrations on the picosecond timescale.⁵ We do not fully understand how EAR achieves such unusually high resolution, but we suspect it is related to the technique's measurement of temperature increase and not coherent scattering, and thus not as sensitive to energy redistribution.

In follow-up work, we measured the vibrational modes of individual fragments of single-stranded DNA with lengths of 20–40 bases.⁶ The DNA acts like a vibrating 1D lattice. Consequently, using 1D lattice theory with no fitting parameters (mass and spring constant values taken from past literature), we were able to match the vibration frequencies observed in experiments to the different lengths and sequences of DNA. With improved resolution in EAR, it may be possible to resolve the base sequence, since each base has a different mass and this will change the overall vibration spectrum.

Moving forward, we aim to apply EAR to the study of protein interactions and viruses. It is interesting to see how potential drug candidates interact with proteins by looking for changes in the characteristic vibrational frequencies. This may provide information about the mechanisms by which drugs function. For viruses, we can use the vibration spectra for detection, to observe heterogeneity in the virus population, and even to learn which resonant frequencies can selectively destroy viruses of interest.